The rigid docking in this case could be unreliable, but a general answer is not possible.

Is the loop under observation known to be extremely flexible? If so you should consider to obtain more conformations of the loop (e.g. with Molecular Dynamics+ Clusterization, or NMR experiments).

Otherwise you can consider to change you approach... for example with a metadynamics flexible docking.

Dear Lesitha,

Is this loop is flexible in nature? In case of some linear and small protein it's quite difficult to perform modeling studies. Still you can go for docking if any co-crystal structure is reported (if not then you can generate multiple grids based on key residues and grid size and can compere data to refine your result).

@Girdhar. Thank you. Actually I have done docking using varied grid dimensions. I wanted to know if there are any report on such kinda predictions. Are there any special way to treat loops. The loop forms a hairpin motif.

I think that in Your case it's better to perform docking, take a few different (the best score and a few that differ in place of interaction) positions of ligand and than perform some Molecular Dynamics simulation (better using full water model). Thus You'll see possible moves of your protein and its answer to interaction with ligand.

The rigid docking in this case could be unreliable, but a general answer is not possible.

Is the loop under observation known to be extremely flexible? If so you should consider to obtain more conformations of the loop (e.g. with Molecular Dynamics+ Clusterization, or NMR experiments).

Otherwise you can consider to change you approach... for example with a metadynamics flexible docking.

I agree with the suggestions of Andrea Coletta.

You should do MD of the protein then Clusterization and choose snapshots to perform docking after protein gets stabilized.

Dear Lasitha,

Loops (Intrinsically disordered structures) have been reported to form ordered structured (binding sites) when in the presence of their binding partners.

So, yes loops, as said in this paper "Structural Ordering of Disordered Ligand-Binding Loops of Biotin Protein Ligase into Active Conformations as a Consequence of Dehydration", in these words " The ligand-binding loops in all known apo BirAs to date are disordered and attain an ordered structure only after undergoing a conformational change upon ligand-binding.", are capable of binding. My work saw this as well.

So, given the correct protocol, you should be able to predict docking reliably.

The question is about adapting the correct protocol.

One way is to find a solved structure X-Ray or NMR. Should you not have that - you will have to model it. The modelling approach is wrong as it introduces errors at a very early stage in your calculations. The X-Ray or NMR should be preferred. Since you already have a protein, i am assuming the structure isn't a problem.

At this point you need to keep in mind probably that loops are disordered, and hence possess a very high degree of flexibility. Therefore any protocol that doesn't take into account this flexibility would not show correct results. You should try Dynamics on this system since they include the flexibility of the site. If you already have a crystal structure as your starting point that is in complex with a ligand and you are substituting one ligand with another and doing comparisons, simple energy minimization routines should serve well, if you keep everything in your system or at least the 5-7 Ang area around the binding site flexible you should obtain reliable results. But you should keep in mind that simple docking is going to give you crude answers. You need to try MCSS (multiple copy simultaeous search) for better results and include solvent effects for still better results. This all depends on the outcome of your investigation.

Hope this helps.

Best,

/A

Dear Lesitha,

You should look at protein as part of a system and not isolated in the gas phase. In this context, you should have some knowledge about the structural and functional behavior of that region? How flexible is that loop? What are the b-factors of that region if the structure is known? Since you are talking about glycine, it is generally a very flexible amino acid and since it does not have side chains deriving specificity is very difficult. Having said that, if your surface pocket is deep, you can still do the docking.

@ Volodymyr, Andrea, Parameswaran. Anthony: Thank you all for the answers. @ Andrea, Ashar. Thanks for the detailed answer. Will apply your suggestions.

As suggested by many researchers the MD simulations is a best way to get a stable confirmation of a flexible protein/loop for any further study.

In addition, repeat the MD several times (at least 2) to obtain better sampling and/or use some better sampling MD approach.

To my understanding, dockings on surface loop are not very reliable. You are never able to reproduce your results. One of my colleagues at University of California, Riverside was working on such system but he failed miserably.

You may think to perform multiple short MD on your protein and then by PCA or normal mode analysis to capture the most relevant movement of your loop. Then you can use the pdb which are capturing those principal movements as inputs in your docking. Or you may perform elastic network model (i.e. good server is http://www.igs.cnrs-mrs.fr/elnemo/)

If an X-ray structure is available, the stability of the functional loop on the surface can be judged by B-factors. If the loop is not stable docking results may fail. If stable MD followed by flexible docking as suggested by many may work.